Trumpism, climate and COVID: Social bases of the new science rejection

statements that are undoubtedly true, such as the Earth is billions of years old, or revolves around the Sun. Conspiracy responses in the June/July and September/October survey stages were nearly identical. There also were no significant shifts in Trump approval, nor in the proportions rejecting ACC or COVID-19. Vaccination eligibility in the US was rolled out by age group in 2021, so older respondents would have had more time to obtain vaccinations, especially in the survey’s June/July stage. To minimize age bias our coding contrasts those who, at the time of the survey, said they were either fully vaccinated, partly vaccinated, or planning to be vacci- nated—versus rejectionists who said they do not plan to be vaccinated. The rejectionist fraction showed only a slight, nonsignificant drop from summer to early fall. Because response patterns stayed nearly the same, we pool summer and fall responses for Table 1 and the analyses that follow. 4.2 Analytical methods Among the endogenous variables listed in Table 1, approval of former president Trump (Trumpism) ranges from 1 (strongly disapprove) to 7 (strongly approve). Most respondents leaned toward one extreme or the other, leaving few in the middle. An alternative dichotomous coding for this question was considered but proved to work less well, as described in section 5.3. Two other endogenous items, climate change and vaccination, offered non-ordinal response choices. Among these choices, rejecting the scientific consensus on ACC, and reject-

ing COVID-19 vaccination, have notably adverse consequences—delaying climate-change mitigation efforts, or worsening the pandemic. Both positions have been linked to conservative identity, including Trump support. For analysis here, those responses are represented by the variables NoACC (reject anthropogenic climate change) and NoVax (reject COVID-19 vaccination). Among the exogenous variables, age is measured in years. Income, education, ideology and party are ordinal items that tend to have monotonic and roughly linear effects in large samples. Section 5.3 evaluates alternative specifications using categorical education and party, but finds that these fit less well, so both are treated as approximately continuous for our main analysis. Codes for education, ideology and party have been centered on zero, for use with interaction effects. Indicator variables for race (white) and religion (Christian evangelical) flag two other factors widely described as correlates of Trump approval and hypothesized to have similar effects on rejection of anthropogenic climate change and COVID-19 vaccination. Gender is treated dichotomously here, although our survey question was more nuanced. Ten out of 1,134 respondents identified their gender as nonbinary, and six others declined to answer this question. These subsets are potentially interesting, but too small for interpretable analysis. Conspiracist ideation has become prominent in US political discourse. The most wide- spread false conspiracy beliefs relate directly to our variables of interest, such as assertions that

Trump really won the 2020 election, climate change is a hoax, or COVID-19 vaccinations con- tain tracking microchips. Obviously, such conspiracy beliefs correlate with other views on Trump, climate or vaccines. The conspiracist variable defined in Table 1 is a less tautological indicator based on receptivity to two extreme beliefs that, unlike Trump, climate or vaccines, have not been explicitly linked to conservative identity: that NASA faked the Moon landings, or that the Earth is flat rather than round. Few respondents in our survey agreed with both views, but more than a third agreed with or were open to at least one. Trumpism is a comparatively recent and still-evolving phenomenon, unlike party, ideology and other demographic characteristics that have been studied through five decades of “social bases” research [9,85,86]. At Trump rallies in particular, Trumpism manifests as a personality PLOS ONE Trumpism, climate and COVID PLOS ONE | https://doi.org/10.1371/journal.pone.0293059 January 10, 2024 6 / 20